https://api.breeze.blue. All public endpoints sit under /v1/* and return JSON or binary audio.
Base URL
Authentication
Most endpoints accept an API key via thexi-api-key header. Account management endpoints such as /v1/api-keys require a browser session instead. See authentication for details.
Content type
JSON requests usecontent-type: application/json. Voice cloning uses multipart/form-data. Audio responses are raw binary.
Errors
Errors returnok: false, a code string, a detail message, and an error field mirroring detail. Validation errors may include a meta object. See errors.
Rate limits
Breeze applies plan-based concurrent generation limits. See rate limits.Explore the API
Convert text to speech
Generate a complete audio response from one text input and saved voice.
Stream text to speech
Receive audio chunks as synthesis progresses for lower-latency playback.
Voices
Understand saved voices, designed voices, cloned voices, and
voice_settings.CLI Reference
Use CLI commands to prototype, inspect API requests, and automate workflows.

